Hyper-Realistic Facial IK Generation Project

I am looking for a Computer Vision and Generative AI expert to create a set of high-precision visual assets for an academic study on facial recognition systems.
The goal is to generate hyper-realistic video and image material from provided reference photographs (ID card and Portrait).
Required Deliverables:
Document Image Enhancement: Process a provided ID card (DNI) photo. This requires correcting perspective (warping), removing light reflections via AI inpainting, and performing 4K/HD upscaling (Super-Resolution) to ensure total data legibility (OCR-ready).
Face Re-enactment Video (10-15s): Create a high-fidelity video using models such as DeepFaceLab, DeepFaceLive, or advanced LivePortrait (local workstation version). The subject must perform natural movements: looking forward, realistic blinking, a slight smile, and a slow head turn (left/right).
Micro-Gesture Variations (2 videos of 5s each): Short clips with subtle eyebrow movements and eye gaze shifts for "liveness" detection testing.
Mandatory Technical Specifications:
Resolution: Native 1920x1080 (Full HD).
Frame Rate: Constant 30 fps (No variable frame rate).
Visual Quality: Total consistency in skin texture (pores, facial hair) and natural cornea reflections. Zero video artifacts (glitches) at the face/neck edges.
Post-processing: Apply a subtle digital sensor noise (ISO grain) layer to simulate an authentic smartphone camera capture.
Budget: $200 USD (Milestone released after technical quality verification).